EPF vs PPF vs NPS: Three Schemes on the Same Criteria

The three differ on who contributes, who chooses, whether the value moves with markets, who carries the risk, what happens at exit and how the arrangement ends. No difference among them can be ranked. The criterion most readers arrive wanting compared, what each will produce, is not knowable for any of the three.

Three names get set against each other constantly: the Employees' Provident Fund (EPF), the Public Provident Fund (PPF) and the National Pension System (NPS). Each of the three is taken apart in full under its own name. The comparison below puts all three on the same set of criteria, one criterion at a time, and reports what each answers.

A comparison is only worth reading if the criteria are ones that can actually be checked, and on this subject exactly one criterion cannot be checked by anybody at all. The uncheckable criterion happens to be the one almost everybody arrives wanting. Six criteria remain after it, and every one of the six has an answer anybody can confirm.

Every rupee below belongs to one invented household carried through this material from the beginning. The Bhosale household is Meghna Bhosale, salaried at Sahyadri Freight Services Private Limited and 36 at the end of the second year, Ashok Bhosale, who runs a tailoring counter, and Ira Bhosale, who is at school. Rs 42,770/- leaves it in an ordinary month, and it holds Rs 3,67,887/- against Rs 71,594/- owed, leaving Rs 2,96,293/-. Two of the three arrangements are in that household and the third is not. Holding two and missing one turns out to teach more than holding all three would.

What are the three arrangements, in one line each?

All three sides need defining before anything is contrasted. A comparison in which one side is described carefully and the other two are left as impressions is an argument with a favourite. One line each is enough to hold the three apart while the criteria run.

The first is a workplace provident fund, an account attached to a job. Money goes into it from two places, from the salary and from the employer, and the scheme announces what it credits on the balance. Nobody in the household is asked to make any decision about it at any point.

The second is a public provident fund, an account a household opens for itself, with the government on the other side of it, into which the household deposits whatever it decides to deposit. The scheme announces what it credits, and the money is held for a term the scheme fixes.

The third is the national pension system, an individual account whose money is put into funds. The account is worth whatever those holdings are worth on the day. The system asks the person whose account it is to make choices, and at exit its rules require part of the balance to be turned into an income rather than handed over as cash.

The same seven questions now go to each of the three in turn.

Why compare on structure rather than on what each will produce?

One question has to be settled before anything else. The answer decides what kind of comparison is even possible.

There are two kinds of thing arrangements can be compared on. A structural criterionSomething about how an arrangement works, checkable today. Who pays in, who decides, what happens at the end: these can be looked up and confirmed by anybody, this afternoon. is something about how the arrangement works: who pays into it, who decides anything, what happens when it ends. The answer can be found today, it is the same answer for everybody asking, and nothing about the future has to be assumed to get it. An outcome criterionWhat an arrangement will produce. It is not checkable today for any arrangement whose result depends on decisions and conditions that have not happened yet. is what an arrangement will produce. Nothing that has not happened yet can be looked up.

Everyday version, and it is worth holding on to. Two bus routes run from one street. Which route starts near the door, who pays for the ticket, who picks where it stops, and where it ends up can all be checked this afternoon, by walking to the stop and reading the board. Nobody at the stop can check which route will get a passenger there faster next Tuesday. The traffic that decides it has not formed yet. A comparison built on last Tuesday's journey times is the past dressed as advice about the future.

Six of the seven criteria here are structural, so anybody can confirm all three answers today. The seventh is an outcome criterion, and nobody can confirm any of the three answers ever. The limit is a property of the subject, and it holds equally for the most expensive comparison anybody has ever produced on it.

Two of these arrangements announce what they credit, and what they announce changes. The third holds units in funds, so it is worth whatever those funds turn out to be worth. Put those together and the outcome column has three cells, every one of them unknown, and three unknowns are not comparable however tidily they are set out.

Criterion one: who actually puts money in?

Start with the plainest question of the seven, because it is the one where the three separate without any interpretation at all. Money arrives in an account from somewhere. Whose money is it, and how many parties are there?

All three are contributoryBuilt from money paid in rather than from a promise. Nobody has undertaken to pay a stated amount later; there is a balance, and it is whatever the money that went in has become.. Being contributory separates all three from the older kind of employer pension. Nobody has promised anybody an amount. There is a balance, built from money that went in.

The workplace one has two contributors. Here, Rs 3,120/- comes out of Meghna Bhosale's salary each month and Rs 3,120/- is added by Sahyadri Freight Services Private Limited: Rs 6,240/- a month going in, or Rs 74,880/- across a year. The employer's side of the two has a name worth knowing, an employer contributionAn amount the employer adds, which is part of the arrangement rather than a gift. It is a term of the job in the same way the salary is, and it does not depend on anybody remembering to ask for it., and it is part of the job in the same way the salary is rather than a favour anybody is doing. The rates behind both figures are set by scheme rules and by statute, and both change.

The public one has one contributor. Meghna Bhosale deposits Rs 1,000/- a month into it, Rs 12,000/- across a year, out of money that has already reached the household. There is no second party. Nobody adds anything alongside her.

The pension system account has one contributor in the ordinary case, the person whose account it is. An employer side exists in some arrangements and not in others, and the rules settle which those are and on what terms. The structural point survives that variation: this is an account belonging to a person rather than to an employment.

Only one of the three brings a second party's money with it, and that one is attached to a formal job. A formal job is a minority position in this country rather than the ordinary one. Ashok Bhosale, at his tailoring counter, has no employer, so no second contributor is available to him at any price. The absence is not a gap in this household's planning. It is what self-employment means, and half the working population is in the same position.

Criterion two: who chooses, and what are they choosing?

The second criterion is where the three separate most sharply, and it is the single most useful of the seven. Asked as a count: how many decisions does each arrangement hand the household?

The workplace one asks nothing. No manager is chosen, no split across kinds of holding, no option selected, no form filled in after the first. Meghna Bhosale has held it for eleven years and has never made a decision about it. There is no decision available to make. It runs whether or not anybody ever looks at it.

The public one asks two things, and only two. How much goes in, and on what date. She answered the first with Rs 1,000/- a month. The second matters more than it sounds. How the credit on this kind of account is computed makes the date of a deposit a real variable, and the computation is set out under the public provident fund. Nothing else is handed to her: no manager to pick, no split to decide.

The pension system account asks three things. Which manager holds the money. How the money is split across kinds of holding. And a third that is easy to miss because it does not arrive as a question at all: whether to choose, or to let a default stand. Leaving the default in place has consequences in exactly the way selecting something does. The third choice is therefore a decision whether or not anybody registers making it.

Everyday version. Three canteens on one street. At the first, a plate is put in front of the diner and there is nothing to decide. At the second, the diner says how much rice and what time, and the rest is settled. At the third the kitchen is handed over: the diner picks the cook, picks what goes on the plate, and decides whether to do either or accept what the cook picks by default. None of the three canteens is the good one. Which is better depends entirely on whether the diner knows how to cook, how much time there is, and whether deciding is a pleasure or a burden. Better belongs to the diner rather than to the canteen.

Criterion three: does the value move with markets?

Now the criterion most people think they are asking about when they ask about risk, and it is worth separating carefully from the one that follows it.

Two of the three carry a declared creditA return the scheme announces rather than produces. The number is set by the party running the scheme and applied to the balance; it does not arrive from anything being bought or sold.. The scheme announces what it credits on a balance, and that announcement is what appears on the statement. Nothing was bought or sold to produce it. Open the statement on a Tuesday when something dramatic has happened somewhere and the balance is exactly what it was on Monday.

The third is market-linkedWhere the value moves with what holdings are worth. The account holds units in funds, and what those units are worth on any given day is what somebody would pay for what sits inside them.. The account holds units in funds, the funds hold things, and the account is worth what those things are worth on the day it is looked at. The value goes up, it goes down, and it does both without anybody doing anything wrong.

A comparison is most often smuggled in at this criterion. A declared credit is not a promise about the future and it is not the absence of risk. A declared credit removes one specific thing: the balance moving because of what somebody else paid for something. Nothing else is removed. The announced number itself changes over time, and each change is set out under the two arrangements that declare one.

A steady number on a statement is a fact about how the number is produced and not a fact about what the money will buy, and confusing those two is the most common error on this entire subject. Prices rise whether or not a statement wobbles. A balance that never moved a rupee out of place across twenty four years has still lost ground to the price of everything, quietly, without one alarming line appearing anywhere.

Criterion four: who carries the risk?

Now the criterion that sounds like the last one and is not. The third question was whether the value moves. The fourth asks who is left holding the consequences, and the answer is a surprise.

Three different risks are hiding inside the single word risk, and separating them is most of the work. The first is that the value moves, and most people mean only this one. The second is that what the money buys keeps shrinking while the balance sits there. The third is that the party on the other side does not do what the arrangement says.

Now run all three arrangements against all three risks. On the first, the two declared-credit ones do not expose the household to a moving value and the market-linked one does. Only that part of this criterion goes the way most people expect. On the second, all three expose it completely. None of the three is a promise about what anything will cost in twenty four years. On the third, all three have a party on the other side; who that party is differs, and the existence of one does not.

Two of the three replace one risk with another and the third carries all of them, so not one of the three takes risk off the household, and the honest answer to who carries the risk is the same in every column. The market-linked one is routinely described as the risky one, and the description quietly implies the other two are not. The three are exposed differently. None of the three is unexposed.

An everyday version makes the second risk visible. The second risk is the one that hides. A household keeps Rs 20,000/- in a tin at home for four years and takes out exactly Rs 20,000/- at the end. Nothing moved, nothing was lost, nobody defaulted, and the tin did precisely what it promised. And the sack of rice that Rs 20,000/- would have bought on the day it went in now costs more. The tin carried no risk of moving and every rupee of the risk that matters over a long enough stretch.

Criterion five: what happens at exit?

Five criteria in, the picture is mostly about how money goes in. The fifth is about how it comes out, and the three separate here in a way that has consequences nobody feels for decades.

The exit ruleWhat must happen to the balance when the arrangement ends. Every arrangement has one; it is settled by the scheme rather than by the household, and it is written down before anybody joins. of the workplace arrangement turns the balance into money under the scheme's own rules. The scheme's rules set when that can happen and on what terms, and they change. The structural point is that what comes out is money.

The exit rule of the public arrangement also turns the balance into money, at the end of a term the scheme fixes. The scheme fixes that length and can change it. Again, what comes out is money.

The exit rule of the pension system account is different in kind rather than in detail. Its rules require part of the balance to be turned into an income rather than handed over as cash. The scheme sets what part, at what age and on what conditions, and every one of those can change. The shape survives every one of those variations: at one of the three exits, a balance the household has watched for decades stops being a balance and becomes an income, and that conversion is a purchase whether or not anybody uses the word on the day.

The distinction between an income somebody promised and an income somebody sold is what this exit rule touches. A household can arrive at the moment of purchase without ever having walked into anything to buy. The household joined an arrangement decades earlier, and the conversion was inside the arrangement the whole time.

A required conversion is not a defect. An income that cannot run out is a real thing to hold and nothing else in household finance produces one. The criterion establishes only that the three exits are not the same exit, and that the difference was settled long before anybody reached it.

Criterion six: how does each arrangement end, and when?

The sixth is the last of the structural criteria and the one readers most often skip. Skipping it is a shame: it decides whether an arrangement survives the things that actually happen to working lives.

The workplace arrangement is attached to an employment. The arrangement came into existence because a job did, and its record follows employments. The scheme rules settle what happens to it when a job ends, changes or breaks, and those rules are set out under the workplace provident fund itself. The structural fact is the attachment: no employment, no arrangement.

The public arrangement is attached to a term. The scheme fixes a length and the arrangement runs to the end of it, whether or not anybody has a job at any point along the way, and whether or not there was ever an employer at all. The term is the whole reason the public arrangement is available to Ashok Bhosale and the workplace one is not.

The pension system account is attached to a person. The account follows the individual rather than a job or a calendar, to an exit its rules define. Somebody who changes work six times keeps the same account.

Each of the three ends differently, and in every one of the three cases the ending was settled on the day the arrangement began rather than decided later by anybody. Reading the ending at the beginning costs an afternoon. Reading it at the end costs whatever it costs.

Here is the everyday version, and it is a wedding hall. Three halls, three bookings. The first is held only while a cousin still works there, so it lasts as long as that job does. The second is booked for a fixed number of years and stands whatever happens to anybody's job. The third is booked in the household's own name and travels with it to another town. Nobody reads those terms on the day. The only thing anybody wants to know then is whether the hall is free, and every one of those endings was decided in the same five minutes.

The seventh criterion: what will each one produce?

Here is the criterion almost everybody arrives with, and it is the reason a table of three figures will not do.

Ask what each of the three will produce over twenty four years and the honest answer, for all three, is that nobody knows. The answer is not modesty and it is not a legal formula at the end of an advertisement. It is a description of what each arrangement actually is. Two of them credit a number the scheme announces, and what gets announced changes. The third holds units in funds and produces whatever those funds produce.

So the seventh column has three cells and every one reads unknown. Three unknowns cannot be ranked, ordered or scored, and no amount of careful presentation turns them into something that can. Three unknowns are not comparableAble to be set side by side on the same basis. Two things are comparable when the same question returns a real answer for both; three unknowns are not made comparable by being printed in three columns. at all.

Which means something specific about every table ever printed with these three names across the top and three figures underneath. Something has been substituted. Either last year's declared numbers stand in for two of the columns, or an assumption stands in for all three. Last year's numbers are a record of the past presented as a property of the arrangement. An assumption is somebody's chosen number wearing the clothes of a finding. Both may be done in complete good faith. Neither turns the unknowable into the known.

Any table showing three figures on this criterion has replaced the thing being compared with something else, and the replacement is invisible precisely because the table looks exactly like a table of facts. The three structural columns beside it may be perfectly accurate. The seventh column borrows the credibility of those six, and the borrowing is what makes it so effective.

A reading that only exists inside a panel is invisible to anybody who has not moved the control. The whole of it is set out below: seven criteria down the left, three arrangements across the top, and the same answers the panel gives at every setting.

Where does this household actually stand across the three?

Six criteria have been run in the abstract. Now put one real position against them. A comparison that never lands on anybody's actual holdings has taught a shape and not a reading.

At the end of the second year the Bhosale household holds two of the three. The workplace arrangement holds Rs 4,12,000/-, built over eleven years of service at Sahyadri Freight Services Private Limited, at a salary that was lower for most of those years. The public arrangement holds Rs 84,000/-, built from Rs 1,000/- a month together with what the scheme has credited on it. There is no pension system account, and Ashok Bhosale has none of the three. One of the three needs an employer, and the other two have never been opened.

Add the two together and the household holds Rs 4,96,000/- against this goal. Of that, Rs 4,12,000/- sits in the workplace arrangement and Rs 84,000/- in the public one. The workplace arrangement has never asked the household a single question. The public one asked two. The workplace share is roughly 83 rupees in every 100.

Run the same reading on what goes in each month rather than on what has accumulated. Rs 6,240/- a month goes into the workplace arrangement, or Rs 74,880/- a year, and not one rupee of it required a decision from anybody in the house. Rs 1,000/- a month goes into the public one, or Rs 12,000/- a year, and every rupee of that was decided. Rs 7,240/- a month in all, Rs 86,880/- a year. For every rupee this household decides to put towards its longest goal, Rs 6.24/- goes in without a decision being taken at all.

The proportion is a fact about how these arrangements are built, not a comment on anybody's attention. The workplace arrangement accumulates precisely because it asks nothing: no remembering, no feeling like it in a difficult month, no deciding on a Sunday evening. Asking nothing is the design. A household in which most of the retirement money came from the arrangement it never decided anything about is one in which an arrangement did what it was built to do.

The other side of the same fact applies to one of the two adults here. An arrangement that accumulates without being decided about is only available to somebody it attaches to. Ashok Bhosale gets nothing from it, and the routes open to him are the ones needing a decision every month, in a house where Rs 42,770/- already leaves each month and the buffer covers 0.73 months. The asymmetry sits inside one household, and it is not a difference in effort between the two adults.

How does somebody assessing a household read these three?

The distinction is not academic, and the quickest way to see that is to watch what different readers do with the same three arrangements. None of these readers is a professional the household has hired. Two of them are strangers and the third is the household itself.

A lender assessing a household reads for two things and runs all three arrangements against both. The first is whether an instalment can be met on the fifth of every month, and that is about income rather than balances. The second is whether a shock can be absorbed without the instalment failing, and that is about what can actually be reached. On that second reading, a balance inside any of these three is a strange object: real, on a statement, and reachable only by scheme rules rather than by the household deciding to. A household that counts an unreachable balance in its buffer has answered a question about its net worth while believing it answered one about its resilience.

A payroll office reads the same three from the other side of the table. For the workplace arrangement it is a remitter: it deducts a side from the salary, adds a side of its own, and sends both. For the public arrangement it has no role and never sees it. For the pension system account it may or may not have a role. Asking which is exactly what to do before assuming anybody is adding anything alongside the household.

And the household reads itself. Reading itself needs no vocabulary at all. Four questions, the same four every time. Who else is paying into this alongside me, if anybody. Which decisions am I being asked to make, and have I made them or let something stand. When can this money be reached, and by what rule rather than by what I would prefer. And what happens when it ends, settled before I joined. A household that can answer those four questions about everything it holds needs nobody's figures at all.

One more reader is worth naming, the one sitting across a table from somebody describing an arrangement. Two questions cut through most of it. Is there a second party paying in, or is it only me. And what does this ask me to decide, now and later. Neither has a number in it, both have plain answers, and a conversation that has not produced them has described a balance and nothing else.

Which question about the three has no answer, and why?

A comparison that quietly avoids the ranking is doing something different from one that states why it will not rank. The refusal is worth saying out loud.

Which of the three is better has no general answer. Not because ranking would be uncomfortable, and not because the rules are complicated. Better depends entirely on the household's position, and the household's position is a set of facts no comparison holds and none can hold.

The facts are ordinary, and they decide everything. Whether there is an employer at all. The answer settles whether one of the three is available at any price. Whether the money will need to be reachable. For a household whose buffer covers 0.73 months that is a different question from what it is with six months in hand. Whether anybody in the house wants to make choices. Wanting to choose is a real preference and not a failing either way: some find a default a relief and others an abdication. How long the goal is. Twenty four years and four are not the same problem. And what else is already held. An arrangement is joined to a position rather than to a blank slate.

Changing any one of those changes the ranking. Dependence on the household's position is not a hedge but the finding itself. A comparison that produced a single ordering would be claiming to know facts about a household that it has no way of knowing, and the more confidently it stated the ordering the less it could have known.

Six criteria can be checked and the seventh cannot. The choice belongs to the household and stays there.

A goal this far out gets read last, if at all, and the reason is worth naming. Richard Thaler and the behavioural economists who followed him documented that people discount distant outcomes far more steeply than near ones: a cost twenty four years away barely registers against a bill due on Friday. Steep discounting is not a character flaw and not something anybody can try harder out of. The only reliable answer is to put the distant thing in front of the household on a day when nothing is due. An afternoon with six criteria is exactly that day.

What is the shortest version of the whole comparison?

Three arrangements, seven questions, six answers each.

On who contributes, one has two parties and two have one. On who chooses, one asks nothing, one asks two things and one asks three. On whether the value moves, two announce a credit and one holds units at market. On who carries the risk, all three answer the household, with what it carries differing underneath. On exit, two produce money and one requires part of the balance to become an income. On how it ends, one is attached to an employment, one to a term and one to a person. On what each will produce, none has an answer and none ever will.

The three differ mainly in how much they hand the household to decide, and knowing where each one sits on that axis is worth more than any figure anybody could put beside them. Further along that axis is not better and is not worse. Further along is only further along.

For the Bhosale household at the end of its second year, the reading is short. The household holds Rs 4,12,000/- in the one that asked nothing and Rs 84,000/- in the one that asked two things, Rs 4,96,000/- in all. Roughly 83 rupees in every 100 of that came from the arrangement it never decided anything about, and that is the arrangement working as built. No account of the third kind is held, and opening one would bring three questions with it. And Ashok Bhosale has none of the three. A tailoring counter with no employer means exactly that, and it is the ordinary position of most self-employed people here rather than a lapse by anybody.

The Bhosale household, Meghna Bhosale, Ashok Bhosale, Ira Bhosale and Sahyadri Freight Services Private Limited are invented.
Educational material. Not advice on any investment, tax, budget or market position.